基于非线性光谱混合模型的遥感影像提取城市不透水层

针对传统方法在提取城市不透水层中的许多局限性,采用两种非线性光谱混合分解模型,包括混合调谐匹配滤波和多层感知器神经网络,通过混合像元分解获取城市不透水层.混合调谐匹配滤波利用用户选择的端元,通过最大化端元响应并减少未知背景信息的影响,进行局部分解端元.多层感知器由多个感知器组成,能够很好的进行非线性学习.对Landsa...

Urban Impervious Surface Extraction from Remote Sensing Image Based on Nonlinear Spectral Mixture Model

Linear Spectral Mixture Model (LSMM) has been used to urban impervious area extraction effectively in recent years. But the assumption of LSMM is usually false, so it is necessary to introduce novel nonlinear spectral mixture algorithms and compare their performance. In this study, three typical linear and nonlinear spectral mixture models were used to decompose the pixels on the remote sensing image to derive urban impervious area information. The fraction images were derived to represent the abundance of four endmembers: vegetation, high-albedo objects, low-albedo objects and soil. Impervious surface was estimated by analyzing high-albedo and low-albedo fraction images. QuickBird multi-spectral image was used to evaluate the accuracy of impervious surface extraction by different methods. Experimental results indicate that the accuracy of artificial neural network is higher than others, so it conclude that non-linear spectral mixture models is also effective to impervious area extraction, even outperform than linear models.